[author]Weidong JI
[content]
China’s Judicial Modernization Driven by Data and Algorithm
Weidong JI
Abstract: This article examines digitalization as a new path for modernization and its operational mechanism, especially the four integrated platforms of smart courts in China today, by taking the vicious circle of “hyper litigant-principle and hyper authority-principle” that had emerged in the transformation of China's legal system as a starting point, taking the games between Algorithm Leviathan and different kind of Digital Monsters as the problem situation, and taking the concept calculation formula and data calculation formula of justice as two basic points. Based on the in-depth analysis of phenomena such as empowering individual judges with artificial intelligence, automatically identifying and pushing similar cases through big data analysis, visualizing trial scenarios, cross domain information sharing through the net check-control system, as well as target management and indicator evaluation in judicial administration, the author elucidates the positive role of data and algorithms in the implementation of modern rule of law principles such as trial independence, due process, judicial openness, and case handling efficiency, as well as various potential risks. Starting from the reality of human-machine coexistence and diverse norms, the author particularly emphasizes the importance of different combinations of legal and technical procedures and their optimization
1.The Modernization and Digitalization of China's Rule of Law
1.1The Goals and Universal Characteristics of Legal and Social Modernization
At the Third Plenary Session of the 18th Central Committee of the Communist Party of China (CPC) in November 2013, the major proposition of "modernize the state system and capacity for governance " was put forward. Since the issuance of the Resolution of the Central Committee of the Communist Party of China on the Major Achievements and Historical Experience of the Party over the Past Century in November 2021, " Chinese path to modernization " has become our country’s underlying values of society. Accordingly, it is necessary to deeply examine and calmly analyze the commonalities of the global modernization and the individuality of modernization achievements in different countries, and then explore functionally equivalent alternatives or substitute options with local characteristics that contribute to the realization of universal goals.
Generally speaking, the remarkable global modernization movement consists of six fundamental elements: (1) the development of machinery and its operational technology, and materialized modernity (Marx's proposition); (2) the progress of natural sciences and the spread of rational thinking (Weber's proposition); (3) the division of labor, cooperation, and scale effects of the market economy resulting from the industrial revolution (Durkheim's proposition); (4) the increase in individual freedom represented by the shift from status to contract (Maine's proposition); (5) the development of the rule of law and social systems engineering (Pound's proposition), and the disciplining of individual behavior (Foucault's proposition); (6) the promotion of subject consciousness and self-discipline (Kant's proposition).
These factors reflect the universal attributes of modernization, but they appeared successively in different historical processes rather than all at once. They also contain contradictions and conflicts with one another. For instance, there is a great tension between the system of absolute monarchy and laissez-faire markets, as well as between individual liberation and discipline. Therefore, we should not understand the commonalities of modern society through a static and absolutely unified framework.
According to the comprehensive research of contemporary American scholar Marc Galanter, by the 20th century, industrial society and modern rule of law presented eleven fundamental characteristics across three major categories: uniform and unvarying, transactional, universalistic, hierarchical, organized bureaucratically, rational, professionals, technical and complex, amendable, political, and differentiated. This constitutes the basic content of the "Law and Modernization" research movement within legal sociology.
For my perspective, however, these descriptions are still too abstract and difficult to grasp or apply accurately in the practice of constructing a modern rule of law order. In fact, the degree of legal modernization can be specifically measured based on eleven indicators across the following four dimensions.
From a constitutional perspective, there are three components at different levels: (1) the concept of popular sovereignty at the top level, (2) checks and balances and local autonomy at the intermediate level, and (3) individual freedom and equality at the grassroots level (which can be measured by the plaintiff's success rate in administrative lawsuits). In the civil law dimension, two key aspects stand out: (4) the protection of property rights and (5) freedom of contract (which can be measured by the enthusiasm for filing civil lawsuits for torts). In the dimension of criminal adjudication, procedural justice is critically constituted by the principles of (6) the legality of crimes and punishments, (7) the non-retroactivity of law, (8) the right to defense and silence, and (9) the openness of trials. In the dimension of civil procedure, the principle of autonomy of will, premised on procedural justice, is reflected in both (10) sufficient communication between parties and (11) the judge’s free discretion.
It should be acknowledged that these eleven indicators represent the main goals in promoting the modernization of the rule of law and judicial reform in China, even if there will be manifestations with local characteristics or innovative measures in specific practices.
1.2.The "Cyclical Phenomenon" of over inquisitorial system and over adversarial system in China’s Judicial Reform
In the process of modernizing China's judicial system, particularly during judicial reforms, a “cyclical phenomenon” emerged—one in which over Inquisitorial system (strong centralized power) and over adversarial system (extensive mediation) exist in a mutually reinforcing yet counterproductive relationship, creating a persistent impasse. It is well-known that Chinese traditional society is characterized by a dense network of interpersonal relationships that form complex, intersecting interactions and private dealings, which often hinder the effective implementation of legal norms.
Such a networked society typically places great emphasis on special interpersonal bonds and reciprocal agreements or specific contracts reached through bargaining. As a result, there is a tendency toward partyism and a self-organized order based on social exchanges. On the other hand, the complexity of the network and fluid reciprocity can hinder the implementation of state directives, making it difficult for policy goals to be fully realized. This necessitates the strengthening of centralized power, inevitably fostering authoritarian tendencies and a managerial order.
Within this context, the formalization, professionalization, and broader modernization of the judiciary tend to advance along a narrow path, focusing on fact-finding efficiency and the enhancement of authoritarianism. It is no surprise that discussions on the rule of law and judicial modernization in China often lead to a technical, scientific approach that emphasizes professionalism, ultimately fostering a movement that introduces and strengthens inquisitorial elements.
As a result, the direct relationship between modernization and authoritarianism is a unique feature of China’s judicial reforms. Particularly when reflecting on the shortcomings of over-partyism and the prioritization of mediation, reforms in adjudication often lean towards over-authoritarianism, emphasizing "adjudication" and neglecting the importance of legal procedures and reasoning in safeguarding individual rights. This, in turn, leads to an expansion of judicial discretion, which, without reasonable institutional checks, is prone to abuse. In this sense, unchecked discretion becomes a catalyst for corruption, undermining the credibility and legitimacy of the judiciary.
To prevent the abuse of discretion and restore the judiciary's reputation, one of the primary strategies for China’s legal modernization has been to strengthen public oversight or to increase the rate of cases resolved through mediation, though there remains no objective, stable institutional framework to serve as a regulator.
Alternatively, another response to China’s judicial modernization has been the complete denial of discretion, such as the experimentation with computerized sentencing. In other words, China’s judicial reforms have followed a logic where discretion is either entirely eliminated through mechanistic measures like computerized sentencing to ensure legal compliance, or it is constrained by public opinion or party recognition, only to reemerge in another form. Breaking this cyclical phenomenon requires a reassessment and redesign of the judicial system and the entire legal order. It is against this backdrop that political and legal institutions have implemented policies such as “People’s Justice,” the “Grand Mediation System,” “Active Judiciary,” and the “Three Key Tasks.”
By rationally analyzing the logic behind this cyclical phenomenon, it becomes evident that the introduction of the "grand mediation" approach was intended to provide legal flexibility in responding to societal uncertainty. The greater the uncertainty, the more pronounced the limitations of professional techniques become, and the stronger the tendency for people to imitate and conform. Particularly in times of intensified social contradictions during a period of transformation, new problems continually emerge, and the boundaries between legality and illegality become blurred. Thus, making clear, black-and-white judgments based on established norms becomes difficult, and mediation and compromise help alleviate the awkward position of adjudicators.
Furthermore, the slogan of “active judiciary” was introduced to address the tendency to avoid litigation and seek remedies outside the system. For years, “people’s justice” has been evaluated by public satisfaction, which means that legal application can be adjusted based on public attitudes, granting courts considerable discretion. Such active judicial policies may foster negotiations that revolve around judicial discretion, strengthening the interaction between courts and the public.
In this interaction, courts adopt the “grand mediation” approach, focusing on resolving disputes through party reconciliation. On the one hand, this encourages litigation, while on the other hand, it maintains the traditional method of dispute resolution through compromise, creating a certain contradiction. Courts, by their nature, cannot satisfy all parties, and in fact, most cases are likely to leave half of the parties (those who lose) dissatisfied. By setting the goal for the judiciary to please everyone, it effectively “ties a noose around its own neck.” This leads to a paradoxical situation where the court, as the final line of defense for state order, is pushed to the forefront of handling social conflicts, placed at the center of public attention, yet judicial authority continues to decline.
1.3.Information and Communication Technology: A New Path for China's Legal Modernization
Since 2010, the rapid advancement of digital information and communication technology (ICT) and its widespread application have presented China with unique opportunities for legal modernization, particularly in judicial reform. The significant assertion that "without informatization, there is no modernization" and "informatization drives modernization" truly encapsulates the essential characteristics of Chinese-style legal modernization.
Theoretically, science and technology have always been crucial indicators of modernization. A historical review reveals that from Leibniz’s mathematical reinterpretation of Roman law to Bentham’s development of the felicific calculus, the essence of legal modernization in Western Europe lies in conceptual computation. Modern jurisprudence seeks to translate all facts into forms that meet legal standards, using axiomatic systems, inclusion relationships, and logical reasoning to mitigate judicial subjectivity, thereby ensuring objective and fair legal judgments and the measurability of markets and society.
Conceptual computation can be summarized by the formula: "Judgment = Formal Logical Syllogism + Subsumption," where the criteria for judicial justice are reasonableness and legality. Since the late 19th century, this line of thought has evolved through various intermediaries, including legal mathematics, jurimetrics, Roscoe Pound’s legal social engineering concepts, jurimetrics, field studies, social experiments, legal and case information retrieval, big data modeling in law, and code framework design in cyberspace, culminating in a highly advanced technological environment.
Notably, Jurimetrics, based on case data analysis, emerged in the late 1940s and gained significant recognition in the early 1960s. This field focuses on the quantitative analysis of judicial behavior through statistical data, modeling judicial processes, assessing the reasonableness and legality of judgments, and measuring the probative value of evidence. This shift signifies a transition in modern judicial systems from conceptual computation to data computation, driven by digital information technology.
Stephen Toulmin's legal argument model provides a direct opportunity for this transition. Toulmin differentiates between claims (C) and facts or data (D) in litigation, further incorporating warrants (W), credibility (Q), rebuttals (R), and backing (B) for argumentation. This creates a complex, data-driven reasoning model that more effectively reflects the needs and operational mechanisms of legal practice compared to traditional term logic structures. This model significantly influences the development of reasoning algorithms in judicial artificial intelligence and digital jurisprudence. When expressed through data computation, the formula can be simplified to "Judgment = Normative Case Big Data + AI Algorithms," where the standard for judicial justice is based on precedents and similar cases.
A key characteristic of "Industry Revolution 4.0" is the interplay of information and communication technology, creating a complementary relationship among the Internet of Things, big data, and artificial intelligence. Network interactions continuously generate big data; AI's analysis and application of this data yield substantial economic value. The larger and higher-quality the data, the stronger AI's performance, creating a feedback loop that forms a digital ecosystem where information and entities converge. This growth spiral increasingly envelops society in digitalization and serves as a mechanism driving continuous legal and social change. It is in this context that the modernization of the data and algorithm-driven judicial system becomes viable.
In China, the issue of arbitrary judicial actions has significantly eroded trust in human judgment, fostering public expectations for the objectivity, neutrality, and certainty of computer-aided sentencing. Consequently, the Internet, big data, and artificial intelligence have rapidly found extensive applications in the Chinese judiciary, leading to a rare phenomenon known as the "smart court boom."
In essence, AI systems that assist in case handling are rule-embedded, which can lead to the rigidity of legal norms and the coding of technology. By ensuring traceability through computer systems, these tools prevent external interference in judicial decisions, thereby reinforcing checks on state power.
In this sense, AI service systems—such as parcel tracking, ride-hailing apps, and traffic violation monitoring—contribute to compliance management and the promotion of modern legal principles. Moreover, the data algorithms and technical frameworks of judicial intelligence can act as leverage, helping judicial activities break free from traditional relational entanglements and power dynamics, facilitating China's judicial reform to escape the "two super paradoxes."
Thus, the court system after 2010 has prominently showcased a modernization of judicial systems and capabilities driven by digital information technology. This transformation has been marked by a series of distinctive phenomena and policy initiatives, including "Judiciary + Data," "Judiciary + Technology," "Judiciary + Ecology," case-based recommendations, automatic document generation, judgment factories, debtor profiling, and joint punishment mechanisms. Ultimately, this evolution necessitates a reevaluation and reinterpretation of legal innovation trends and essences in the digital age through the lens of the interplay among pre-modernity, modernity, and post-modernity.
1.4.The Need for Paradigm Innovation in Legal Studies in the Digital Age
The modernization of the judicial system and capabilities driven by data and algorithms hinges on the digitalization of the judicial field. Key milestones in China's judicial digitalization process include the implementation of the first five-year outline of the program for reform of people’s courts in 1999, the comprehensive adoption of information and communication technology (ICT) and equipment, and the promotion of case flow management systems. Beginning in 2010, judicial rulings from various courts gradually became available online; in 2016, the establishment of smart courts and internet courts commenced; by 2019, there was a top-down emphasis on paperless court operations; and after 2020, online filing and hearings became widespread. Over the past 25 years, particularly the initiatives of publishing judicial rulings online and transitioning to paperless court operations have significantly advanced the digitalization of case handling, achieving a considerable degree of digital coverage in the judicial field.
In 2023, the Central Committee of the Communist Party of China and the State Council jointly issued the " Plan for the Overall Layout of Building a Digital China," aiming to promote Chinese-style modernization through dual tracks of digital infrastructure and data resource systems, with the goal of establishing a clear competitive advantage by 2035. It is important to note that the interplay between digitalization and modernization offers significant opportunities and conditions for building a rule-of-law society in China and advancing judicial reform, but it also presents numerous challenges and risks.
To prevent the runaway of the "algorithmic Leviathan," can we adopt a "digital constitutionalism" or "digital rule of law" similar to how Enlightenment thinkers and politicians since the late 17th century strengthened institutional guarantees for individual freedoms and rights, thus safeguarding "digital human rights"? Alternatively, can we leverage the existing power of digital platform giants (for example, GAFA in the U.S. — Google, Apple, Facebook, Amazon; or BATH in China — Baidu, Alibaba, Tencent, Huawei), along with sovereign individuals active in blockchain, as well as the combined strengths of large language models and multimodal models to counterbalance the algorithmic Leviathan, creating a cyberspace reminiscent of the feudal system of medieval Western Europe or the decentralized power checks of bureaucratic "indirect governance" in China?
In my view, this multi-layered and diverse digital power structure could also be termed "digital relationalism," or we might explore a third path that combines these two prospects, tailoring it to the actual situation in China. For instance, implementing a normative dual-track system during the transformation to a digital society. This would involve enhancing the protection of digital human rights through fair processes and sufficient communication, based on the mutual checks between Leviathan and digital monsters.
From the perspective of "digital relationalism," the rise of digital platform giants in the Internet 2.0 era, the formation of numerous virtual communities in the Internet 3.0 era, and the emergence of various universal models in the AI 4.0 era all constitute intermediary groups within the digital stratification of society. These intermediary groups can interact and integrate beyond the borders of sovereign states, exerting global influence, while also establishing autonomous rules and mechanisms for dispute resolution outside of national jurisdiction. For example, the internal regulations of digital platforms and virtual communities may include technical standards, algorithms, and codes, which can be enforced through various sanctions such as points, upgrades, service preferences, material rewards, tax incentives, and the allocation of voting rights, as well as through restrictions like banning participation and speech, removing web pages, or suspending accounts. Additionally, they may connect with real-world legal systems through purpose trust agreements, legislative negotiations, and internet courts.
In this intricate web of interactions, institutional design must shift its focus from outcomes to processes, from prior administrative approvals to post facto judicial remedies, and from unidimensional evolution to multidimensional interaction. To prevent the abuse of power by large platforms and models through embedded fair processes in AI systems, as well as to counterbalance various algorithmic powers (including those of AI assistants), the principles of procedural fairness in modern rule of law and judicial authority should be redefined and play a more significant role.
Since algorithms constitute an important form of power in the digital age, there is a need for legitimacy arguments and normative constraints regarding algorithms. These arguments and constraints often hinge on principles of procedural fairness, encompassing both technical and legal dimensions, necessitating the adoption of typified approaches to technical procedural fairness and legal procedural fairness. The essence of technical procedural fairness is to safeguard individuals' digital rights through code regulation. If such regulation takes a legal form, then the intersection and combination of technical procedural fairness and legal procedural fairness can be identified. In a sense, finding the optimal combination of these two aspects represents a key legal connotation for the continued advancement of modernization in the digital state, which can be understood as a starting point for digital constitutionalism or digital rule of law, and undoubtedly serves as a prerequisite for innovation in adjudication paradigms.
In addition to ensuring that argumentative dialogues occur smoothly under conditions of freedom and equality, attention must also be paid to the technical procedural fairness of data processing, algorithm design, and execution. This involves addressing a series of procedural appropriateness questions, such as how codes and their frameworks should be properly regulated, who the authors of the code are, who can control the authors, whether there are prior validation procedures for the appropriateness of the code or corrective procedures afterward, how the powers and responsibilities of internet service providers as intermediaries should be allocated, what regulations apply to software development involving data flows, and whether digital surveillance and internet searches are subject to restrictive conditions.
2. Transformation of the Trial Paradigm Driven by Data and Algorithms
In specific judicial contexts, the modernization of court functions through electronic computer technology has long been a topic of academic discussion. However, implementing data and algorithm-driven judicial reform comprehensively from both value and practical perspectives to form a national system is an unprecedented experience for China. In April 2020, the Supreme People's Court issued the "Opinions on Implementing the Spirit of the Fourth Plenary Session of the 19th CPC Central Committee to Promote the Modernization of the Trial System and Trial Capabilities" (Fa Fa [2020] No. 9), which explicitly states the need to fully utilize the functions of the people's courts, accelerate the modernization of the trial system and trial capabilities, and effectively enhance national governance efficiency. It specifically mentions the need to "integrate trial execution work deeply with modern technology, and deepen the construction and application of smart courts," indicating the foundational institutional construction role that big data and artificial intelligence play in advancing the modernization of the trial system and trial capabilities.
The core goal of modernizing the trial system and trial capabilities is to ensure the smooth operation of trial authority and better demonstrate the principle of independently exercise adjudicatory power by people's courts, thereby more fully meeting the fundamental concepts of constructing a modern rule of law order, which is to better reflect the normative connotation of trial authority. For example, this includes strictly handling cases according to the law, preventing subjective arbitrariness and abuse of discretion by judges, strengthening the rationality of trial forms and procedural fairness, improving the efficiency of judicial system operations, increasing the predictability and legitimacy of judgments, etc. These are also the main functions of applying new technologies like the internet, big data, cloud computing, blockchain, and artificial intelligence in judicial contexts.
Here, we first need to understand the institutional logic of data-driven judicial reform and the construction of smart courts; only then can we truly realize the functions and benefits of digital technology in the modernization of the trial system and trial capabilities. At the same time, we must fully recognize the various profound impacts of big data and artificial intelligence on the mechanisms of judicial authority, especially the associated risks, side effects, and potential harms. Therefore, when discussing data and algorithm-driven judicial reform and the modernization of the rule of law, we also need to emphasize data governance and algorithm governance, ensuring that artificial intelligence is controllable and trustworthy, and that the entire rule of law system and society are safe.
2.1Institutional Logic of Data-Driven Judicial Reform and Smart Court Construction
Contemporary China's modernization of the rule of law faces a dilemma: the lack of Public trust accountability credibility. How to enhance judicial credibility is a pressing issue that needs to be addressed and is a fundamental goal of court system reform. In summary, the components of judicial credibility primarily include the following four aspects.
1. Judicial impartiality: The Constitution of our country stipulates the principle of independently exercise adjudicatory power and prosecutorial power. The reform and opening up have prompted institutional changes, with the basic trend shifting from prior administrative licensing and direct regulation to post-judicial relief and indirect guidance. Therefore, the courts must maintain a neutral stance that transcends different interest groups and possess the ultimate decision-making power over legal issues, which inevitably raises the technical standards of the court. Here, the key to independently exercising trial authority is to limit the abuse of discretion by judges, prevent individual case errors or even wrongful judgments, and ensure the objective operation of the legal system based on uniform judgments in similar cases, equal status among parties, and impartial fairness and justice.
This relies on several supporting conditions, such as the design of fair procedures and reason-giving activities conducted through legal interpretation, reasoning, and more open discussions. The distinct feature of China's modernization of the rule of law is the emphasis on accountability measures for wrongful judgments while establishing the principle of independently exercising trial authority, ensuring judicial fairness by holding judges accountable for consequences. This underscores the importance of the judicial responsibility system in the modernization of the Chinese trial system. Another way to achieve judicial fairness is through intelligent analysis using big data to better realize judicial justice in similar cases.
2. Judicial authority: Selecting excellent talents who are both morally and artistically upright through very strict qualification criteria to become judges, empowering courts with the authority for legal judgments, ensuring that effective judgments possess immutable res judicata and unwavering enforceability, and instituting constitutional reviews of administrative and legislative powers all contribute to enhancing judicial authority. Due to the judicial supervision process and petitioning system, Chinese court decisions lack finality and res judicata; constitutional reviews are limited to specific administrative actions, which means that judicial authority largely relies on traditional "Imperial Examination Faith," reflected in the need to pass rigorous judicial and civil service examinations to become a judge. In the judicial reforms of the 2010s, a selection process for quota judges was added, making the judicial quota system an important initiative for reshaping judicial authority in contemporary China in the sense of meritocracy. Simultaneously, faith in science and the efficiency and precision of technology also help establish judicial authority.
3. Judicial efficiency: If cases are delayed, justice is always late, leading to public distrust in the judiciary; thus, judicial efficiency is vital for judicial credibility. During periods of social transformation, various conflicts arise, and economic relationships become increasingly complex, resulting in a surge in litigation cases and a high annual average of cases handled per judge, necessitating a significant increase in the efficiency of hearing work. The widespread application of big data and artificial intelligence clearly benefits the operational efficiency of the judicial mechanism and facilitates timely judicial relief, improving the satisfaction of parties and stakeholders.
4. Judicial accessibility: The trial system must facilitate access for parties; otherwise, the public will distance itself from the system. Experts in civil procedure law and comparative law, such as Capelleti, focus on "Access to Justice," with a key demand being to improve the "hard to enter" situation of courts. In 20th-century China, the mass line of the judiciary was a primary method for facilitating judicial accessibility, openness, and participation, but information and communication technology clearly provides faster and more effective means. Powerful legal information search engines, the increasingly popular online case filing and hearing systems embed a consumer-oriented service concept into the judicial system, promoting judicial transparency and participation.
Against this backdrop and under these conditions, to break the "dual super-circle" and enhance judicial credibility, China initiated a new round of judicial reforms in 2016. This included measures such as "delocalization," "de-administration," "Two systems" process management, "One Planning plus Three Reviews," "tracking and trackability" (limiting batch approval trials), and "intelligence" (limiting rough trials). The focus has been on advancing the judicial responsibility system, judicial quota system, and judicial intelligence attempts, aiming to leverage these reforms for the modernization of the trial system and trial capabilities. It is evident that the internet, big data, and artificial intelligence play crucial roles in this process.
2.2.The Impact of Big Data and Artificial Intelligence on the Operation Mechanism of Judicial Power
Current practices demonstrate that the Internet of Things, big data, cloud computing, and artificial intelligence have significantly enhanced the efficiency and quality of the judiciary, while also effectively improving judicial services. Thus, it is essential to organically embed digital technology within judicial system reforms, emphasizing the roles of data-driven approaches and algorithmic support in the modernization of judicial systems and capabilities. Strengthening the development of intelligent courts aligns with the necessary establishment of the rule of law in a digitally transformed society. While it is crucial to support such reforms to break free from the “double overshoot trap,” we must also maintain a calm and thorough understanding of the potential impacts, risks, threats, side effects, and various chain reactions brought by new technologies, and implement necessary and robust measures for data compliance and AI governance.
2.1.1Challenges to Existing Modern Rule of Law Principles
Artificial intelligence fundamentally operates as an electronic computing system for processing data. It generates feature quantities from data and models phenomena, executing tasks of matching, calculating, predicting, generating, and decision-making through algorithms between inputs and outputs. With the advancement of machine learning, the trend toward algorithmic opacity is becoming increasingly pronounced, leading to varying degrees of autonomy in AI systems. In this context, probability-based predictive models can easily foster a results-oriented technological mindset, undermining process-oriented judicial thinking. For instance, both the “Predictive Policing System” and the “System of Joint Punishment for Losing Faith” exhibit a tendency to prioritize results over procedural fairness.
The so-called “Predictive Policing System” involves utilizing AI to analyze big data and perform probability calculations to capture the spatiotemporal dynamics of criminal activities, thereby predicting behaviors and trends of unlawful acts. Undoubtedly, predictive policing aids in the investigation and proof of criminal offenses and helps prevent urban security risks. The 5G communication system expands the application of big data and AI, making technology installation and operation more convenient, thus enhancing predictive policing capabilities. This type of predictive policing system and prosecutorial framework is not only widespread in China but also utilized in Europe (e.g., the CAS in Amsterdam, Netherlands), the U.S. (e.g., PredPol in California), and other Asian countries (e.g., CRIMENABI in Kyoto, Japan). However, such systems do prompt significant changes in the institutional arrangements of modern rule of law.
First, the predictive basis of policing relies on historical data and social conditions, necessitating the processing of big data according to specific standards to differentiate and classify individuals, essentially labeling citizens. It is evident that this process of classification, labeling, and targeted monitoring can easily result in algorithmic discrimination.
Second, AI systems utilize algorithms to conduct probability calculations on big data, inferring that certain individuals possess a higher propensity for criminal behavior, which inevitably leads to stricter monitoring of these individuals, effectively applying different legal treatments and contravening the principle of equality before the law.
Finally, it is crucial to note that modern rule-of-law nations emphasize the presumption of innocence in criminal matters to prevent wrongful accusations. However, predictive policing and related prosecutorial systems, by preemptively implementing monitoring measures against specific groups, may lead to a “suspicion breeds guilt” effect, causing individuals to be wrongfully accused based on labels. In a sense, this can be seen as operating on a presumption of guilt. Therefore, stricter inspections, audits, validations, and evaluations of such systems from the perspectives of data ethics and algorithmic fairness are necessary.
The “System of Joint Punishment for Losing Faith” is more familiar to the Chinese public. In recent years, relevant Chinese authorities have been promoting credit scoring practices and joint credit punishment systems to enhance social creditworthiness. Notably, the Supreme People's Court has signed cooperation agreements with Alibaba’s Sesame Credit scoring system, as well as with various institutions in asset management, financial services, and transportation, aiming to improve debt recovery and judgment enforcement through scoring systems and joint punishment mechanisms.
For a long time, the judiciary faced significant challenges regarding enforcement issues: from 1985 to 1987, the execution rate of economic dispute rulings in Chinese courts was only 20%-30%, and by the early 1990s, the execution rate for civil and commercial cases was still only 30%-40%. Leveraging intelligent analysis of big data to monitor the property transfers and movements of defendants, combined with coordinated processing of big data from other domains, could significantly enhance the effectiveness of judicial measures such as litigation preservation, debt recovery, and judgment enforcement.
Conversely, from the perspective of law and justice, scoring systems like Sesame Credit also present various problems and risks. A notable concern is that these systems classify and label individuals based on consumption big data, leading to differentiated treatment in areas such as school admissions, loans, car rentals, hotel stays, insurance purchases, and international travel, thereby creating de facto discrimination and violating the constitutional principle of equality before the law. In summary, while credit evaluation systems provide convenient tools for law enforcement, judicial actions, and social governance, they also introduce new challenges that require strengthened risk prevention and regulatory measures
2.2.2.The Potential Alienation of the Modern Judicial System Due to Big Data Processing
It is important to note that the application of big data and artificial intelligence in the judiciary can lead to certain paradoxes. Smart courts were established to improve judicial efficiency, uphold procedural fairness, and realize modern rule-of-law ideals. However, the use of digital technologies may also distort the judicial mechanism. For instance, the draft judicial opinions generated by AI systems are automatically produced based on existing judicial big data, where intrinsic systemic biases from past regulations and cases can become entrenched or even amplified in the big data processing. In other words, while data and algorithms can correct individual biases, they struggle to rectify systemic biases.
Under these circumstances, the opportunities for policy and normative innovations during case hearings, as well as the gradual promotion of institutional changes through the accumulation of individual cases and their chain reactions, may be undermined, resulting in new dilemmas for judicial reform. Even if it is emphasized that judicial AI is merely an auxiliary tool and judges are the primary decision-makers, the heavy caseloads faced by judges and the strict enforcement of accountability for erroneous judgments can easily promote a bias towards automated judicial decision-making. This outcome not only contradicts the principles of AI governance but also exacerbates the impact of systemic biases.
Moreover, the widespread application of big data and artificial intelligence may lead to the hollowing out or even disappearance of crucial legal interpretations, reasoning, and discourse that are vital to the judiciary. In modern rule-of-law nations, the key to enhancing judicial credibility lies in procedural fairness and thorough argumentative dialogue and communication conducted under this premise.
If the operation of smart courts does not rely on research into legal reasoning algorithms and related software design, and if the relationship between legal interpretation and intelligent adjudication is not adequately considered, failing to make thorough dialogue between parties a necessary component of the process, it could undermine or negate the principle of direct speech in modern adjudication. As a result, judicial processes risk being reduced to a formula of "AI + arbitration," losing the rich meanings and justification mechanisms generated from interactions, while the institutional arrangements for tiered adjudication and the division of legal professions become dispensable.
In handling cases, judges form intuitive judgments and internal convictions based on the factual circumstances, understanding of normative content, value judgments, rules of thumb, balancing interests, sense of fairness, and common sense—this is referred to as the principle of free conviction. Naturally, judges verify, substantiate, revise, or negate this conviction based on evidence and argumentative activities, ultimately leading to a formal judicial decision. If AI systems are to make judicial determinations, it would require transforming all relevant information—including laws, precedents, case details, experience, feelings, perceptions, and common sense—into a computer-readable format, necessitating the construction of a vast and comprehensive knowledge base and knowledge graph.
It is evident that building such an extensive knowledge base is an endless endeavor. This reflects what John McCarthy referred to as the "frame problem" in his 1969 theories of artificial intelligence. Furthermore, humans possess "tacit knowledge" that cannot be expressed in concrete terms, which is crucial for intuitive judgment; yet AI struggles to comprehend and process this tacit knowledge that cannot be converted into symbols and algorithms, a phenomenon known as the "Polanyi Paradox" in AI theory.
Encouragingly, the emergence of language models and multimodal models since late 2022, employing techniques such as data compression, distillation technology, prompt-based word prediction, and the integration of deductive and inductive reasoning, seemingly provides a feasible path to address the frame problem and the Polanyi paradox.
Judicial decisions often involve value judgments, cultural characteristics, and linguistic meanings, necessitating the establishment of relationships between the real world and the world of meanings. Understanding linguistic meanings is particularly crucial in the case hearing process; traditional Chinese practices like the "Five Sounds for Judging Cases" and the emphasis on "hearing disputes" illustrate this point. While AI systems can convert language into conceptual definitions and computational rules, they cannot genuinely grasp the emotions and meanings inherent in language; to AI systems, any emotionally laden discourse is merely a computational symbol, and the value embedded in those symbols cannot be realized or experienced.
Thus, Claude E. Shannon’s information theory conceptualizes information using the physical concept of "entropy" and abstract mathematical forms, detaching it from meaning to build probabilistic language models that can generate language. However, the probabilistic realm cannot encapsulate the meaning of the world. This reflects what Stefan Hanard referred to as the "symbol grounding problem" in AI theory in 1990. Even with significant breakthroughs in generative AI's ability to simulate human thought and form, it still lacks the self-awareness that humans possess due to emotions and cannot build social connections or accumulate social capital in the same way. In this context, the demand for emotional value and meaning will become increasingly significant, potentially leading to a public choice shift from rational to emotional justice, with people possibly expecting judges to focus more on realizing emotional values.
The protection of a judge's professional identity is a fundamental principle in the institutional design of modern rule-of-law nations. The underlying rationale for this arrangement is that, in order to ensure judicial fairness, judges must undergo the most rigorous selection processes and handle cases within the most transparent judicial proceedings, rendering judgments based on open debate and thorough reasoning. Consequently, a qualified judge should make decisions strictly in accordance with the law, free from external interference. The final judgment document should represent the optimal juncture for resolving all disputes and serve as the last line of defense for social justice. In this sense, it is imperative to guarantee that judges enjoy a certain degree of identity privilege and sufficient subjectivity.
The judicial reforms in China during the 1990s saw attempts to implement a system of appointing presiding judges based on competition and performance-related pay, which tended to undermine the established judicial identity through market-oriented approaches. Following the introduction of artificial intelligence systems into the judiciary, judges' actions are increasingly subject to pervasive surveillance, and judicial reasoning is often influenced by automated document generation technologies, further eroding the already tenuous identity privileges and subjectivity of judges. For instance, while the judicial entity in China's institutional design has never been singular, it has diversified even further in the digital age. In practice, the construction of smart courts has transformed judgments into collaborative decisions made jointly by judges, computer engineers, software designers, and data processing contractors. There is even a potential for algorithmic dictatorship under certain conditions, or for judges to excessively rely on expert system software, thereby diminishing their own subjectivity.
Once artificial intelligence systems are deeply integrated into decision-making, the boundaries of judicial entities and their powers can easily become blurred, complicating the foundations and causal relationships of judgments. This complexity is particularly exacerbated by the black-box nature of algorithms, making it increasingly difficult to ascertain responsibility, thus rendering judicial accountability a mere formality. Therefore, the design of judicial artificial intelligence systems must adhere to a "human-centered" principle, strictly preventing the eventual dominance of algorithms or automated machines over judicial determinations. Additionally, it should recognize, as stipulated in the EU's General Data Protection Regulation, the right of parties to disengage from automated decision-making.
The above encapsulates the primary risks, issues, and potential harms associated with the modernization of a data- and algorithm-driven judicial system. To address these concerns, it is essential to enhance data governance, algorithm governance, application system governance, and large model governance across three dimensions: ethical principles, legal frameworks, and technical standards. This approach will maximize the benefits of digital technology while mitigating various side effects, ensuring that smart courts remain on a trajectory of robust development.
3.Innovations in Modern Judicial Concepts and Mechanisms Achieved Through Four Digital Platforms
Based on primary data and various reports, it is evident that the new round of judicial reforms driven by data and algorithms is primarily manifested in the construction and operation of four integrated platforms: the Smart Trial Platform, the Smart Enforcement Platform, the Smart Litigation Service Platform, and the Smart Management Platform. The following analysis will specifically examine how these digital technologies actualize modern judicial concepts and promote institutional and mechanism innovations.
3.1. Smart Trial Platform
Undoubtedly, the modernization of the judicial system hinges on the principle of independently exercising judicial power, one prerequisite of which is the modernization of judges' capabilities. Historically, a significant barrier to the independent exercise of judicial power has been the uneven professional quality and limited case-handling abilities of judges. However, the Smart Trial Platform, as the core of the judicial system, can indeed empower judges through technologies and methods such as case push notifications, electronic evidence presentation, machine reasoning, expert software, automated document generation, and remote electronic signatures. This empowerment substantially enhances case-handling efficiency and helps prevent the abuse of discretionary power.
Taking the Shanghai Intelligent Auxiliary System for Criminal Case Handling (“Project 206”) as an example, this system enables the entire case process to be managed online and operates under a "single-set" electronic file system. Through comprehensive digital and intelligent applications such as intelligent perception throughout the process, operational traceability, and dynamic monitoring, the system aims to improve judicial efficiency, quality, and transparency. Notably, it significantly enhances the guidelines for evidentiary standards across different litigation stages in criminal cases by leveraging big data and artificial intelligence. This advancement caters to the personalized needs of conviction and sentencing while strengthening logical reasoning and judgment predictions during evidence chain reviews. The use of digital technology to eliminate illegal evidence and reasonable doubts from the evidence chain at different stages is not only a fundamental objective of algorithmic design but also a concrete manifestation of the principle of due process. Thus, procedural control over judicial activities weaves the concepts and elements of modern justice into the technical details of data processing.
On January 23, 2019, the Shanghai Second Intermediate People's Court conducted its first trial using the Intelligent Auxiliary System for Criminal Case Handling, publicly addressing the "Yin Robbery and Murder Case." This marked the system's success in connecting the entire process of criminal cases from filing, investigation, arrest, prosecution, to trial. According to survey statistics, by the end of 2019, the Shanghai Second Intermediate Court and 11 other courts, including those in Xuhui, Fengxian, Railway Transport, Jiading, Qingpu, Chongming, Changning, Jing'an, Minhang, and Huangpu, had held 18 hearings utilizing this AI-assisted trial system. Compared to earlier smart judicial practices, the "Project 206" has achieved (or partially achieved) intelligent review of evidence materials, effectively separating the authentication work related to case facts from judges' adjudication duties, significantly improving case-handling efficiency and aiding precise case assessments. Additionally, the design and application of intelligent auxiliary systems for civil, commercial, and administrative cases are gradually advancing.
It is clear that the key to the success of various intelligent auxiliary case-handling systems lies in how to establish steps and specific criteria for identifying similar cases within vast amounts of data, thereby realizing the goal of just adjudication in the context of similar cases and judgments.
From the practical experiences of courts in Shanghai and elsewhere, it is observed that, on one hand, the Smart Trial Platform aims to leverage information technologies such as voice recognition, image recognition, video retrieval, feature extraction, and semantic understanding to assist judges in verifying evidence. By utilizing a judgment document auxiliary generation module, the system conducts simultaneous checks on the creation of judgments, promptly alerts for defects, and prevents errors, thereby facilitating more convenient extraction and use of electronic file contents. It also promotes intelligent case diversion applications to enhance trial efficiency, using the efficient retrieval capabilities of big data and artificial intelligence to achieve case fairness and judicial justice, thereby reinforcing the realization of parties' rights.
In summary, the utilization of digital information technologies has revolutionized the allocation of roles and resources within the judiciary, inevitably prompting paradigm innovations in adjudication.
Moreover, the comprehensive monitoring of judges’ behaviors may lead to excessive formalization and mechanization of case handling, potentially causing psychological backlash among judges. It is also essential to prevent the phenomenon of "AI judges handling cases entirely" and to implement necessary measures to mitigate the ethical, legal, and accountability risks associated with automated decision-making.
3.2. Smart Enforcement Platform
In 1999, to address the longstanding issue of enforcement difficulties within the judicial system, the Central Committee of the Communist Party of China issued Document No. 11, titled "Notification from the Central Committee of the Communist Party of China on Forwarding the Report from the Leading Party Group of the Supreme People's Court of the Communist Party of China Regarding the Resolution of the 'Difficulties in Enforcement' Issues in the People's Courts." This initiative aimed to mobilize all relevant forces across the country to participate in enforcement, designating that year as the "Year of Enforcement."
However, more than a decade later, the phenomenon of enforcement difficulties remained unresolved. In this digital context, the introduction of the "Network Control System" emerged as a critical tool for precise enforcement based on big data-driven intelligent analysis. According to the Supreme People's Court's "Report on the Status of Judicial Enforcement Difficulties," by September 2018, prior to the pilot launch of comprehensive judicial reforms in Shanghai, courts nationwide had utilized the network control system to provide inquiry and freezing services for 57.46 million cases, freezing a total of 299.2 billion yuan, demonstrating significant effectiveness.
The application of this system varies across different regions, with unique characteristics in each court. For instance, the Beijing court system focuses on creating credit profiles for debtors, the Zhejiang court system emphasizes establishing a case-specific account for each individual case, and the Sichuan court system is experimenting with intelligent pricing for debtor assets. The foundation of the Shanghai court system's Smart Enforcement Platform is also this network control system, which monitors debtor assets throughout the enforcement process by leveraging cross-domain data applications. It effectively enhances the recovery rate of claims and standardizes enforcement procedures through algorithmic support.
Here, the key to improving enforcement efficiency lies in enhancing data exchange between internal and external networks, gradually forming an intelligent command and scheduling system that covers the entire process and responds promptly. By harnessing digital technologies, this system automates the backfilling of case information, formats judgment documents, associates party information, prompts execution milestones, activates network controls automatically, publicizes the enforcement process, pushes enforcement leads automatically, and issues alerts for violations.
This system employs credit scoring and big data analysis to construct profiles of debtors, utilizing the network control module to track the whereabouts of debtors, discover clues related to enforceable assets, and link disciplinary data for those who default on court orders. It also promotes the integration of inquiry assessment data with national assessment industry association databases to enhance parties' bargaining power through the asset evaluation module. Additionally, an online application module for attorney investigation orders expands the scope and avenues for enforcement leads. The fines management module ensures that in cases where effective judgments specify fines, the enforcement department initiates execution and automatically associates enforcement information with the fines management system, thereby improving the implementation rate of monetary penalties.
3.3 Smart Litigation Service Platform
In Shanghai, the Smart Litigation Service Platform consists of several components, including the litigation service hall, litigation service website, mobile litigation service applications, and the "12368" hotline. This platform seeks to provide a combination of online and offline intelligent litigation services, offering streamlined options such as "one-stop service, one-network service, one-number service, and one-time service," all based on the principles of the internet, big data, and cloud computing.
The specific services provided by this system include case registration and filing, assignment and adjustment of trials, assistance in judgment execution, multi-faceted dispute resolution, and handling litigation-related petitions. It enables online consultations, assessments, mediations, arbitrations, lawsuits, and claims, as well as facilitates the assignment of mediators, pushes similar case notifications, displays evidence online, provides online judicial confirmations, and conducts online service of documents. It is particularly noteworthy that, while enhancing convenience for the public, the Smart Litigation Service Platform further promotes judicial transparency by optimizing the openness of trial processes, courtroom activities, judgment documents, and enforcement information through remote operations, efficiency, and synchronous interaction.
According to the "Opinion of the Supreme People's Court on Building a One-Stop Multi-Dispute Resolution Mechanism and One-Stop Litigation Service Center" (Fa Fa [2019] No. 19), released in August 2019, the Smart Litigation Service places greater emphasis on the integration and fusion of platforms. It aims to connect over 3,500 courts nationwide to the China Mobile Micro Court mini-program, interlinking with mediation platforms, preservation platforms, expert assessment platforms, and various court systems for trial processes and litigation fee payments. This integration enables parties to access comprehensive litigation services through a single public service entrance and embeds social governance mechanisms within the framework. Once this convenient infrastructure is established, it will provide essential conditions for enhancing the uniformity of legal implementation and fully leveraging the role of lawyers. Concurrently, issues regarding how to embed self-monitoring and self-correcting mechanisms within artificial systems, as well as how to establish checks and balances among multiple artificial intelligence systems, must also be placed on the agenda for discussion.
4. Smart Litigation Service Platform
In Shanghai, the Smart Litigation Service Platform consists of the litigation service hall, litigation service website, mobile litigation service applications, and the "12368" hotline, aiming to provide parties with integrated online and offline intelligent litigation services based on the Internet, big data, and cloud computing, including forms like "one-stop service, one network for all, one number for all, and one-time handling" (including litigation service apps or WeChat mini-programs).
The specific content of the litigation services provided by the system includes registration and case filing, sorting between complex and simple legal cases, judicial execution assistance, multi-faceted dispute resolution, and litigant-related petitions. It allows for online consultation, online assessment, online mediation, online arbitration, online litigation, and online claims processing. The platform also facilitates the assignment of mediators, case pushing for similar cases, online evidence presentation, online judicial confirmation, and online delivery. It is particularly important to emphasize that while the Smart Litigation Service Platform serves to enhance convenience for the public, it can also further promote judicial transparency by optimizing the public processes of trial, courtroom activities, judgment documents, and execution information through remote operations, efficient handling, and synchronous interaction.
According to the "Opinions on Establishing One-Stop Multidimensional Dispute Resolution Mechanisms and One-Stop Litigation Service Centers" issued by the Supreme People's Court in August 2019 (Fa Fa [2019] No. 19), smart litigation services increasingly focus on the integration and convergence of platforms, ensuring that over 3,500 courts nationwide are fully online with the China Mobile Micro Court mini-program, and are interconnected with mediation platforms, preservation platforms, entrusted appraisal platforms, as well as the litigation processes and litigation fee payment systems of various local courts. This allows parties to receive a "one-stop" litigation service through a single public service entry while embedding social governance mechanisms. Once this convenient infrastructure is established, it will provide the necessary conditions for improving the uniformity of legal enforcement and fully leveraging the role of lawyers. At the same time, it is also essential to address how to embed self-monitoring and self-correction mechanisms within human systems and how to establish checks and balances among multiple artificial intelligence systems.
4.1Smart Management Platform
The main function of the smart court management platform is to promote the rationalization, compliance, standardization, and precision of judicial administration. Utilizing judicial big data analysis systems and information communication technologies, it aims to enhance the capabilities of scientific office operations, document handling, and the efficiency of court operations, achieving judicial transparency in a visual manner. Additionally, local court systems have also researched the "fine calculation" of case trial workloads, categorizing cases based on case types and indicators such as "fixed weight coefficient," "floating weight coefficient," and "general weight coefficient," thereby calculating case workload, evaluating judges' performance, and determining the number of judges.
For instance, the Shanghai First Intermediate People's Court employs the smart management platform to sort between complex and simple legal cases and to differentiate handling, making the transition between single-judge and collegial systems smoother, and allowing for a more equitable distribution of judges' workloads. In practice, the reform of the judgment system driven by data and algorithms has particularly notable effects on improving the efficiency of administrative management and achieving precise judicial goals, resulting in courts resembling "judgment factories."
5.Conclusion and Outlook: Embedding the Principle of Procedural Justice into Data Processing Processes
The traditional Chinese concept of "emotion, reason, law" creates significant discretionary space in judicial and law enforcement activities, and the intertwined interactive relationships can easily undermine the effectiveness and certainty of regulations. Promoting judicial reform in such a context often finds itself trapped in a cycle of indecision between "Hyper-Party Doctrine" and "overstepping authority," making it difficult to break free. Over the past thirty years, the widespread adoption of digital coverage and information communication technology seems to provide a robust leverage to escape this cycle, potentially paving a new path for the modernization of Chinese-style rule of law—achieving judgment fairness in terms of similar cases and judgments while enhancing the procedural controllability and outcome predictability of interactive relationships through code architecture, all while improving the operational efficiency of courts.
Generally speaking, artificial intelligence is fundamentally a system that operates according to rules and codes within established parameters, leading to the rigidification of legal norms and preventing external powers or influences from interfering with judicial judgments through comprehensive tracking by computer systems. This reinforces the constraints on state power. In this sense, it can be argued that analyzing and predicting big data through artificial intelligence systems, along with broader digital processing, promotes the spirit of modern rule of law. Therefore, the advancements in big data and artificial intelligence have ushered the modernization of the Chinese trial system and its capabilities into a new phase, seemingly creating a stunning landscape of technology-driven rule of law.
This round of judicial reform driven by data and algorithms essentially involves embedding the principles of modern adjudication into the data processing information systems and various operational processes of courts at all levels. This allows the objectives required by modern adjudication—objectivity, neutrality, fairness, formal rationality, normative unity, judicial transparency, convenience for litigation, and efficiency in case handling—to be achieved more fully and perfectly through computer science and information technology. In this sweeping institutional transformation, the integration of judiciary and technology creates a digitally covered legal landscape, facilitating precise adjudication, full-process management, and resource integration. Along this logical extension, it becomes evident that big data and artificial intelligence have also fostered an unprecedented ecosystem, altering the social environment surrounding courts and transforming the interactive relationships between public security, prosecution, and judicial agencies as well as lawyers and parties involved. The paradigm of adjudication is undergoing profound changes, necessitating a reevaluation of modernity issues and a careful examination of the latent ethical, legal, and procedural risks.
The rapid development of artificial intelligence in recent years has triggered a significant transformation in industrial models and even social structures. We observe that in 2000, artificial intelligence was still considered a heteronomous system focused on logical calculations. However, after 2012, the emergence and strengthening of deep learning functionalities have gradually transformed artificial intelligence into a self-regulating system. This current state of artificial intelligence, particularly generative artificial intelligence since late 2022, is characterized by the ability to induce from facts and data rather than merely operate according to human-given instructions or programs.
Big data and artificial intelligence are continuously evolving at an accelerated pace and penetrating the judicial fields of various countries. Reports indicate that since early 2023, ChatGPT has been widely used in judicial and legal service scenarios, including handling legal business for the "Big Four" accounting firms globally. On May 5 of the same year, Lexis Nexis, a top global legal information service platform, launched the generative AI system Lexis+AITM for the legal sector, aimed at achieving efficient legal and case retrieval as well as automated document generation through large language models. Although two lawyers in the U.S. were penalized in mid-June for citing false cases provided by ChatGPT, by the end of the year, the UK judiciary issued official guidelines for judges to use large language models, permitting judges to employ ChatGPT in drafting legal documents.
In reality, whether it is language models or multimodal models, at their core, they rely on vast databases and various mathematical formulas. From 2023 to 2024, AI large models have made significant inroads into the judicial sectors of various countries, and the dream of "legal mathematics" seems to be within reach, further demonstrating the significance of big data for the modernization of trial systems and judicial capabilities, while also suggesting that the paradigm of rule of law will continue to innovate and undergo revolutionary changes. Studying this subject holds immense academic value and practical significance.
However, it is crucial to recognize that the genuine application of artificial intelligence in legal decision-making has a long way to go. Artificial intelligence is more adept at processing or learning vast amounts of data according to established patterns, significantly improving the efficiency of primary tasks. Nonetheless, using massive datasets for machine learning poses threats to data security and personal privacy, and it can easily result in algorithmic bias due to systemic inaccuracies in historical data.
In March 2024, the EU issued the "Artificial Intelligence Act," categorizing AI risks into four levels: unacceptable, high, limited, and minimal, each with distinct regulatory approaches. Of particular note, the Act identifies AI applications deemed extremely harmful and contrary to European values, such as manipulating personal behavior through social scoring systems, real-time remote biometric identification technologies, and predictive policing systems, as falling under the prohibition category.
Additionally, legal expert systems assisting judges and lawyers, as well as intelligent trial projects, are also classified as high-risk types that require stringent regulation. This stance may have some impact on the construction of smart courts in China, especially in the context of handling foreign-related cases. In summary, during the development of AI large models, maintaining an appropriate balance between safety and development is the primary challenge facing contemporary artificial intelligence governance and law, and it is also crucial for the construction of smart courts in China.
From the perspective of modernizing trial systems under digital coverage, embedding procedural justice into judicial big data processing involves procedural control of court operations, where the intersection and combination of technical procedural justice and legal procedural justice hold particularly significant implications. Here, technical procedural justice includes aspects such as the coding of legal procedures, process management through technological tools, and the demonstration and auditing of algorithms, all closely related to artificial intelligence governance.
Although there has been some basic consensus in the international community regarding the governance of artificial intelligence, such as the nine key principles of "human-centered," "collaboration," "sharing," "fairness," "transparency," "privacy," "social safety," "technological safety," and "accountability," differing interpretations exist behind similar expressions, and significant variations can be found in the institutional and mechanism designs of artificial intelligence governance. In the realm of judicial artificial intelligence governance, deeper divisions exist in understanding and addressing the balance of authority, procedural justice, ethical considerations, and risks, resulting in considerable challenges in advancing smart courts, particularly in terms of judicial application.
To promote the construction of smart courts with a focus on procedural justice in the age of artificial intelligence, it is essential to develop a systematic understanding and practical path that meets the needs of a new era. Firstly, it is vital to understand the implications of procedural justice deeply, comprehensively exploring the significance of procedural justice to judicial fairness and operational efficiency in the contemporary context. Secondly, it is crucial to recognize the various risks posed by artificial intelligence in judicial applications and to implement targeted regulations. Finally, an in-depth examination of the path toward establishing procedural justice within the framework of artificial intelligence and judicial applications is necessary.